Pragmatic patterns and the lexical system—A reassessment of evaluation in language

During recent years, corpus-based phraseology has brought new insights to the study of language use and in particular to the role of evaluation in language. New quantitative studies (e.g. Stubbs, 2001b) reveal that rather than choosing individual words and their meanings, the average writer/speaker often chooses larger building blocks, so to speak, as the meanings of words are much more intrinsically connected than is traditionally assumed. A given word is statistically likely to trigger other words and frequently also a set of connotations, as evidenced by the concept of semantic prosody (Louw, 1993). From the point of view of meaning, it makes more sense to talk about extended units of meaning (Sinclair, 1996) than about the syntagmatically more limited concept of lexical meaning. This article discusses the implications of these new insights on the semantics/pragmatics boundary, in particular the claim that the traditionally primarily pragmatic concept of evaluation is part of the language system to a much larger degree than hitherto presumed. It offers Danish language analyses which corroborate English and Italian findings within the area of semantic prosody; these analyses are furthermore used to argue for the existence of inherent evaluative meaning and pragmatic evaluative meaning.